Business Opportunities from Generative AI and Agentic Workflows

Agentic AI Potential

In his keynote at BUILD 2024, Andrew Ng, Founder and Executive Chairman of Landing AI, highlighted the transformative potential of generative AI and agentic reasoning. With advancements in large language models (LLMs) and large multi-modal models (LMMs), AI is entering a new phase of development. The focus is shifting toward AI agents and their ability to harness unstructured data—text, images, video, and audio—to build innovative applications across industries.

Generative AI as a Catalyst for Transformation

Generative AI is revolutionizing industries by enabling faster, more efficient workflows and driving innovation at unprecedented speeds. Much like electricity transformed industries a century ago, generative AI is proving to be a general-purpose technology with broad, transformative potential.

One of its most significant contributions lies in accelerating the development of machine learning models. Traditional workflows often involved months of data labeling, model training, and deployment. Generative AI compresses these timelines into days, allowing teams to prototype and test multiple ideas quickly. For example, in retail, generative AI can analyze customer sentiment from thousands of product reviews in mere hours, enabling businesses to adjust marketing strategies or improve product designs in real time.

This rapid pace fosters a culture of experimentation, empowering organizations to adopt a “move fast and be responsible” approach. By reducing the time required to innovate, generative AI encourages teams to explore diverse ideas while maintaining a focus on reliability and accountability. For instance, a healthcare startup might use generative AI to design personalized treatment plans by analyzing patient histories and medical research, significantly reducing the time required to provide actionable insights.

The impact extends beyond speed. Generative AI democratizes technology by making it accessible to a broader audience. Its ability to generate complex outputs from simple prompts enables even non-experts to leverage its capabilities. For example, a small e-commerce business could use generative AI tools to create marketing content, such as personalized ad copy or product descriptions, without needing a dedicated creative team. Similarly, in manufacturing, generative AI can optimize supply chain logistics by predicting demand patterns and identifying cost-saving opportunities, all with minimal human intervention.

The Rise of Agentic AI Workflows

Agentic AI workflows are redefining how we interact with and utilize artificial intelligence. By integrating iterative processes like research, critique, and revision, these workflows empower AI systems to tackle complex tasks more effectively. This shift marks a departure from traditional, linear AI operations toward a dynamic and collaborative approach.

At the core of agentic workflows are four key design patterns: reflection, task planning, multi-agent collaboration, and reasoning. Reflection enables AI models to critique and refine their own outputs, improving quality through continuous self-assessment. For example, in software development, an AI system can write code, evaluate it for bugs or inefficiencies, and iteratively improve the output without requiring significant human oversight.

Task planning allows models to sequence actions and handle intricate requests, such as analyzing a legal contract before drafting a summary with highlighted risks. This capability is particularly beneficial in fields like law or finance, where precision and thoroughness are paramount.

Multi-agent collaboration takes this a step further by enabling different AI models to specialize in distinct tasks. For instance, in healthcare, one AI agent might analyze patient imaging data, while another reviews medical literature to recommend treatment options. Together, these agents provide a comprehensive and well-rounded solution.

Reasoning patterns also enable AI systems to manage complex operations, such as planning a marketing campaign. An AI might identify target demographics, create promotional content, and schedule advertisements based on analytics, all while adapting the strategy dynamically as new data comes in.

Agentic workflows prioritize rapid experimentation without sacrificing reliability. They embody the evolution from “move fast and break things” to “move fast and be responsible,” encouraging innovation while maintaining accountability. For example, in retail, an agentic AI might develop personalized shopping experiences by constantly iterating and refining recommendations based on real-time customer interactions.